Embodiments according to the invention relate to generating virtual images and, in particular, to a device and a method for providing a video signal of a virtual image based on a real image of a camera.
Video data generally only have implicit optical location or position information available, from which only the relation of visible objects to each other may be determined. A determination of a real positional information in real time is only possible via reference models within the image or cost and time consuming image processing. Cross-fading objects in video signals in real time until now necessitated a high computational analysis and processing of the video image, without the possibility to resolve the ambiguity of depth with regard to size independent of the object and doubtlessly.
Real time location systems exist which are able to measure the positions of objects within a certain area to the accuracy of only a few centimeters. Examples are Fraunhofer Witrack, Differential GPS or Ubisense. The position of the objects may be illustrated as a virtual 3D view in the user interface.
Further, there are fully automatically controllable camera systems whose exact alignment and also position may be determined by integrated sensors or odometry. Also the position to be expected may be used, to which the camera is controlled, as far as highly accurate camera heads are used.
Cameras may, for example, also be controlled based on data from a location system.
On the other hand, there are different systems for cross-fading video signals in real time from two or more sources of analog or digital video signals, and the same already enable fading-in additional information (e.g. date, time, running text . . . ) into running video images in real time. Further, cross-fading is possible with predefined graphics and animations. Video processing systems like face recognition already allow fading-in image data supported information. The image processing overhead is very high, however.